Announcements

A new release of SubString is now available.

Highlights of the new release include the ability to consolidate n-gram lists down to the level of 1-grams and various fixes for Linux compatibility.

---

detailed release notes v. 0.9.4

substring.sh:

• consolidation down to single words enabled (before it was down to bigrams)
• fixed some linux-specific issues with coreutil's cut
• fixed an issue where a warning was issued about unexpected format in lists had n-grams with underscores in them
• efficiency improved by reducing the number of I/Os
• improved accuracy during imports at preparatory stage

cutoff.sh:

• improved efficiency
• introduced -i option for improved naming of cutoff lists

length-adjust.sh (new):

• newly added script to adjust lengths of certain n-grams after consolidation

listconv.sh:

• fixed bug in -n option
• expanded number of input list formats handled

test_data:

• updated gold lists to reflect changes to substring.sh
• new data in example3 that includes word lists
• new example4

README.pdf/README.md:

• adjusted for changes
• description for length-adjust.sh added